It's wholly imaginable different governments person been the unfortunate of rogue AI agents.
Former Australian authorities cybersecurity advisor Alastair MacGibbon told the BBC he'd heard whispers that respective others person been notified of akin caller breaches by OpenAI agents.
"Some person chosen to not beryllium nationalist – that's each government's prime connected however it wants to grip these things," helium said. "The [Australian] authorities chose a clip to merchandise this to summation maximum publicity which is their wont to do."
Revealing a information breach tin of people beryllium a risky strategy for governments - it leaves them susceptible to disapproval that their information systems aren't up to scratch. But the information that nary delicate accusation was leaked enactment Australia successful a stronger presumption to usage the incident.
"Nobody has died," says the University of Queensland's subordinate prof Michael Noetel, who studies AI risks. "This is different canary successful the ember mine. This benignant of loss-of-control incident, adjacent though it's insignificant now, is what CEOs are disquieted astir getting worse implicit time."
Though Australia has made a sanction for itself by taking a basal against societal media companies, taking up the AI mantle present is different mode for Australia to rein successful large tech, says Tama Leaver, prof of net studies astatine Curtin University successful Perth.
"It's intolerable to accidental for sure, but it seems incredibly apt that this was precise cautiously planned."
.png)














English (US) ·